Grok Gains Ground
Just weeks after its stock market debut, SpaceXAI has unveiled its latest artificial intelligence model, Grok 4.5.
In a blog post published on Wednesday, SpaceXAI described the new release as a highly capable and dependable "workhorse" model. The company said Grok 4.5 is designed to handle a wide range of everyday knowledge-based tasks that the AI industry seeks to automate, including coding, application development, office work, research, writing and other routine workflows.
According to SpaceXAI, Grok 4.5 is also significantly more cost-efficient to operate. The company claims the model delivers "twice the token efficiency" of many leading competitors. If the claim proves accurate in real-world use, it could provide a major competitive advantage, as rising token costs have become an increasing concern for AI users.
The company also released several benchmark results on Wednesday, showing that Grok 4.5 performs competitively against its leading rivals, although it falls just short of achieving best-in-class status across all benchmarks.
In a post on X, founder Elon Musk compared the new model with Anthropic's Opus model, a large language model specifically designed for highly complex reasoning and in-depth research tasks.
Musk wrote, "Based on the overwhelmingly positive feedback from customers in our beta testing program, SpaceXAI will make Grok 4.5 available to the public on Thursday. It is essentially an Opus-class model, but much faster, far more token-efficient and considerably less expensive."
He later added, "According to our internal evaluations, Grok 4.5 is roughly comparable to Opus 4.7 in capability, while being substantially faster. This combination of strong performance, speed and lower cost makes it highly competitive in the market."
SpaceXAI said the new model is priced at $2 per million input tokens and $6 per million output tokens. If its real-world performance matches the company's claims, the pricing would make Grok 4.5 one of the more cost-effective and competitive models currently available.
By comparison, Anthropic's Opus 4.7 costs $5 per million input tokens and $25 per million output tokens. OpenAI, meanwhile, follows a tiered pricing structure across its model lineup. Its premium Sol model is priced at $5 per million input tokens and $30 per million output tokens, while its entry-level Lunar model costs $1 per million input tokens and $6 per million output tokens.
Competition in the AI model market has intensified this week. OpenAI is also expected to release its latest and most advanced model, GPT-5.6, on Thursday. The model's launch had previously been delayed due to security concerns raised by the Trump administration. OpenAI has described it as its "most powerful model yet."
//DBTech/BMT/OR//
